Hi Yuxuan thanks for updating the MR the overall output looks ok to me the last tiny bit I'm still concerned with is the 6th patch (the one checking if pid is the current process); i theory it shouldn't be harmful, unless we end up with different access rights in toolhelp's openprocess vs the current process pseudo handle ones but it could (theorically) change existing behavior (for a regular module snapshot) considering that: * there's no report (at least that I know of) of an app making use of 32bit module snapshot * the only real client of the API (with 32bits) will be dbghelp so I'd rather not include the 6th patch (until we get evidence we need it), and rather ensures dbghelp passes the 0 pid when dealing with current process (and testing accurate access rights) -- https://gitlab.winehq.org/wine/wine/-/merge_requests/9371#note_127697